Is Evidence for the Laws of Nature Also Evidence Against a Miracle?


In two previous posts on the topic of miracles we learned three important points which are worth reiterating here:

First, a miracle can be defined as a religiously significant event which is caused by God and exceeds the productive powers of nature.

Second, the productive powers of nature can be understood using various theories of natural laws: they can be reduced to law-like generalizations about what repeatedly occurs in nature, expressed in terms of necessary truths about what must happen under natural conditions, or be viewed as descriptions of what physical objects are disposed to produce under certain circumstances. Now, rather than defend a particular understanding of the laws of nature in this post, I will settle for a rather uncontroversial definition proposed by J.L. Mackie (1982): the laws of nature “describe the ways in which the world – including, of course, human beings – works when left to itself, when not interfered with.”[1]

Third, we learned that David Hume (1748) is very skeptical about the possibility of believing in a miracle. He argued that it is never rational to believe in a miracle because the evidence for the laws of nature is always stronger than the evidence for a miraculous exception to those laws.[2] At the very least, Hume assumed the evidence for one counts as evidence against the other. [It is worth noting that Mackie agrees with Hume on this point]

But was Hume correct in his assumption that the evidence for the laws of nature is evidence against a miracle? Richard Otte (1996) disagrees, and the purpose of today’s post is to examine why he does so.

balance of evidence11
Is Evidence for the Laws of Nature Also Evidence Against a Miracle?


Otte claims that the laws of nature tell us how nature works when left to its own devices, when no outside factors like God are acting to exceed its powers. This means that belief in a naturally occurring exception to the laws of nature flies in the face of all the observational evidence that those laws hold true. However, Otte points out that those laws do not tell us how nature must work when outside factors are interfering. Why is this point important? Because that is exactly what is happening when a miracle occurs! A miracle (by definition) only occurs when nature has not been left to itself, when its powers have been exceeded by God. Therefore, evidence for the laws of nature is not evidence against how nature would work if a being like God decided to act in such a way.[3] As Otte (1996) himself puts it, “Evidence for a law of nature is evidence about how the world works when God does not intervene, but a miracle is what happens when God does intervene. Since evidence for a law of nature is not evidence about how the world works when God does intervene, it is not evidence against a miracle occurring” (p.155).[4]

Richard Otte (1996) uses an everyday example from house plumbing to illustrate the fallacy of counting evidence for the laws of nature against belief in miracles.[5] He writes as follows:

Evidence for how my water lines normally behave (without interference) is not evidence against how they’d behave when my plumber is interfering – Richard Otte (1996)

“My experience is that when a plumber doesn’t interfere with my plumbing, whenever I turn on the faucet water comes out. I might justifiably conclude that whenever I turn on the faucet and a plumber doesn’t interfere, water will come out. But it would be a mistake to conclude that when a plumber interferes with my plumbing, water will come out when I turn on the faucet. It is very possible that because of the plumber’s interference, when I turn on the faucet water won’t come out. How my plumbing behaves when the plumber does not interfere with my plumbing does not determine what I should believe about how my plumbing behaves when the plumber does interfere. If Mackie’s reasoning were correct it would be irrational to believe someone who claimed the plumber turned off the water to their home and water didn’t come out of the faucet” (pp.154-155).


Otte’s conclusion is important to the rationality of belief in miracles because it is often claimed that the laws of nature make belief in miracles so antecedently improbable than no amount of historical evidence could ever counterbalance it. But if Otte is correct, then evidence for these laws has no direct bearing on the antecedent probability of a miracle. It only has indirect bearing insofar as we have evidence to think that God would not act to exceed the powers of nature. Such evidence can take the form of arguments against the existence of God or against his having reasons to cause a miracle on some (or any) particular occasion.

[1] J.L. Mackie, The Miracle of Theism Oxford: Clarendon Press, 1982, pp. 19-20.

[2] David Hume, An Enquiry Concerning Human Understanding. 2nd Edition. Ed. Eric Steinberg (Cambridge: Hackett Publishing Company, 1993), pp.72-90.

[3] Richard Otte, “Mackie’s Treatment of Miracles.” International Journal for Philosophy of Religion 39:151-158 (June 1996).

[4] Otte admits that “it is possible that there be evidence that is relevant both to how the world behaves when God intervenes and when God does not intervene” (p.158). But evidence for the laws of nature does not appear to be like this.

[5] Otte borrowed this example from Barbara Scholz.

19 thoughts on “Is Evidence for the Laws of Nature Also Evidence Against a Miracle?”

  1. I don’t see how the concept of evidence can even be applied to miracles.

    Evidence is the effect from which we infer a cause. If we find a body with a knife in its back and the knife has little swirly patterns on the handle, we infer that the person whose fingerprints match those patterns is the person who put the knife there. We can do this because we understand the natural processes of cause and effect that cause those little swirly patterns to appear on objects and we believe that those natural processes are invariable. If we thought that those little swirly patterns appeared randomly or by divine fiat, they wouldn’t be evidence of anything.

    Our method of drawing inferences from evidence depends on the consistency of the natural processes of cause and effect that we observe. That does not prove that those natural processes could never be suspended, but I think we are precluded from inferring that they have been in any particular case. The most we can do is to say that we don’t know what the cause is for a particular effect.

    1. That’s a very good question vinnyjh57, because it brings up questions about the nature of explanation and sources for causal claims.

      It is correct to say that observing regularities is one way to detect causal connections between phenomena, but the claim that this is a *requirement* seems to have fallen on hard times in the philosophy of science. Why? Because many philosophers of science recognize that causal explanation can be had without that requirement.

      First, we sometimes detect causal powers directly without needing a sample of observed regularities in support of a hypothesis. To use an everyday example – I can reasonably conclude that a particular itch on my arm caused me to scratch it, without my needing an inductive sample of itchings and scratchings in support of that conclusion.

      Second, the history of science over the past two hundred years is full of cases where hypotheses postulate ‘theoretical entities’. These entities are hypothesized to possess the causal powers that would be *needed* to increase the likelihood of the phenomena in need of explanation. The causal explanation is not inferred by regularly observing the entities together with their effects, since the entities may not be directly observable, or even observable in principle.

      So in both cases above, causal explanation (and, therefore, the concept of evidence that goes with it) is not restricted to the inductive method you described.

      1. My guess would be that the connection between itches and scratching is a regularity that you have observed pretty often.

        But my point is not that direct observation is the only way to determine cause and effect relationships. The point is that observed cause and effect relationships are based on natural laws that act consistently. That is how we are able to infer a cause from an effect. If we didn’t believe that those natural laws acted consistently, we couldn’t infer from a pile of smoldering ashes that there had been a fire. We would have to allow for the possibility that they had appeared randomly or by divine fiat.

        1. Hi vinnyjh57,

          Thanks for clarifying your position.

          If I understand you correctly, your claim is that we need a world that operates consistently in order make any causal inferences whatsoever, whether through direct observation, induction, postulating theoretical entities, or otherwise. Have I got that right?

          I like your word “consistently” because your first comment used the word “invariable.” My hunch is that invariability is not necessary for the kind of consistency we need. We only need enough consistency to be able to make probabilistic predictions, and our ability to make those predictions would not be compromised if there were infrequent exceptions to the norm.

          How does this relate to natural laws? Two tentative thoughts: first, an exception to a putative law would not automatically invalidate it, or require us to formulate a new law. The old law would only be invalidated or in need of replacement if the exceptional event occurred repeatedly under the same conditions. Second, the purpose of my main post (above) was to argue that the hypothesis of a miracle is not evidence that an exception to a law of nature has occurred. No law is being suspended because laws of nature only specify how nature works when left to itself, and a miracle by definition is an event that occurs when nature is not left to itself.

          Of course, this raises a very difficult philosophical question – namely, what would justify the claim that a particular event was not brought about by a law of nature, but was instead caused by divine agent? My most recent post, entitled “Can Miracles Be Rationally Believed” attempts to lay a foundation for responding to that question. Here is the link:

          In that post, I define the concept of ‘evidence’ broadly enough for it to be used for or against a miracle claim.


  2. Sometimes skeptics argue that there could never be evidence for a miracle, even in principle, because our ability to make causal inferences depends on laws of nature which cannot have exceptions. Why? Because exceptions would undermine our ability to make the predictions on which those inferences are based.

    But this objection to the possibility of evidence for a miracle is flawed on two counts: first, the skeptic cannot simply assume that rare exceptions to the laws of nature would undermine our ability to make predictions. Laws that operate reliably, but with infrequent exceptions, would still give us the predictability we need. Predictability does not require that laws be perfectly consistent, only consistent enough, for causal inferences. Second, the objection begs the question because (as argued above) miracles are not (strictly speaking) exceptions to those laws.


    1. It is not simply that we need the consistency of natural laws to draw inferences. We use that consistency to draw inferences.

      A mercury thermometer both needs and uses the consistency of the relationship of the temperature of mercury and its volume. We might posit the possibility of exceptions to that relationship, but a mercury thermometer will only give us an accurate reading in the unexceptional cases. By the same token, we won’t be able to infer the correct cause in a case where cause and effect are not determined by natural law even if we think such cases might exist.

      1. All this shows is that when an exception to a natural law is occurring, a causal inference (which would otherwise be correct) would be mistaken. But what is this conclusion supposed to show? That we can never know for certain when those inferences are correct and when they are not? This is only a problem if 100% certainty is a requirement for knowledge, and most philosophers reject that requirement as too restrictive.

        Epistemologically speaking, evidence applies to a belief insofar as there can be facts, events, or experiences that increase or decrease the probability of that belief in relation to its competitors. This evidence need not make the belief 100% certain in order to be evidence.

        Likewise, insofar as the thermometer indicates a correct temperature in the vast majority of cases, it makes the correctness of any particular reading highly probable, even if there is a slight probability that it will yield an incorrect reading. Thermometers (and the natural processes they were designed to measure) don’t have to be 100% reliable or consistent in order to increase the probability of an inference based upon them. All we need is probability, not certainty. Therefore, the thermometer still functions as evidence.


        1. The problem isn’t that the possibility of exceptions undermines our ability to make predictions. The problem is that we cannot predict the exception. We cannot use a thermometer to infer a temperature other than the one that is consistent with the normal relationship between temperature and volume even if we can imagine that such a temperature might occur. By the same token, we cannot infer a cause of any particular effect other than one that is consistent with the normal operations of the laws of nature even if we posit that such causes might exist.

  3. We do not have an accurate reading from the thermometer when the law doesn’t hold. So the reading in that specific case would be mistaken. I agree with that.

    But I am claiming this: because the thermometer is accurate in the vast majority of cases, any particular reading has a high probability of being accurate. Would you agree?

  4. My hunch is that we are operating on two distinct understandings of what evidence is. I maintain that E is ‘evidence’ only if inferences based on E are *probably* true. Whereas you seem hold that E is ‘evidence’ only if inferences based on E are true. That distinction might explain some of our disagreement.

    p.s. actually i overstated my position before. Rather, E is ‘evidence’ only if inferences based on E increase in probability with respect to E.

    1. I don’t think that I have a problem with inferences that are probably true.

      What I am arguing is that the justification for drawing inferences in the first place is the existence of known cause and effect relationships. If I didn’t know what caused water to fall from the sky, I couldn’t infer that it is raining from the feeling of water of landing on my head.

    1. Suppose you are sitting on a jury in a murder case where two witnesses claim to have seen the defendant shoot the victim, but ballistics tests show that the bullets that killed the victim came from someone else’s gun. Even if your world view includes a God who could supernaturally alter the firing marks on bullets if it suited His purposes, could you claim that the evidence pointed in that direction?

      1. I suppose not, but the reason would not have to do with the laws of nature, supposing god exists. The reason would have to do with not having adequate reason to think the God of theism would have reason to alter the ballistics in that case.

          1. I don’t presume to know that. But in the example you used, altering the ballistics seems rather arbitrary. When addressing the probability of a miracle, things are different because the religio – historical context would need to be such that the event serves a religious purpose and be the kind of event that God may have reason to bring about. At the very least, for belief in the miracle to be rational, the probability of the event, given God’s existence, cannot be so low that no amount of posterior historical evidence could make the event more likely than not.

          2. You say that you don’t presume know God’s purposes and yet you expect to assess the religious purpose of an event in order to determine whether God might have a reason to bring it about. The only way I can see to determine what God’s purposes might be is by examining the things He has said and done, but it is impossible for me to know whether something was said or done by God without knowing something about His purposes first. As God is certainly under no obligation to make his purposes known to me, it might well be that anything He chooses to do would necessarily appear arbitrary to me.

Leave a Reply

Your email address will not be published. Required fields are marked *